👋 Hi, I’m Andre and welcome to my weekly newsletter, Data-driven VC. Every Tuesday, I publish “Insights” to digest the most relevant startup research & reports, and every Thursday, I publish “Essays” that cover hands-on insights about data-driven innovation & AI in VC. Follow along to understand how startup investing becomes more data-driven, why it matters, and what it means for you.
Current subscribers: 21,255, +205 since last week
Brought to you by Harmonic - Monitoring Your Firm’s Network at Scale
Your firm’s network is one of its most valuable assets – are you leveraging it appropriately? With Harmonic, you can map your team’s Linkedin, email, and calendar connections to our startup database, so you’ll always be the first to know when someone in-network is launching “something new”. Beyond that, our customers have been using connection mapping to:
Find the best path to potential investments
Search for thesis fits that are exclusively in (or outside of) the firm’s shared network
Make introductions to potential customers for portcos
Enrich their CRM with their team’s connections
The venture capital industry has started to massively professionalize in the past years. We have seen various new models evolving such as boutiques vs asset aggregators, specialists vs generalists, early vs growth vs multi-stage firms, solo vs micro GPs, and a lot more.
One dimension that cuts horizontally through all of these different models is digitization and the use of data & AI. It took 70+ years to start moving from the traditional all-human handcraft setup with manual, inefficient, and oftentimes ineffective workflows to a data-driven model where algorithms and automation augment humans who stay in control of the outcome.
Though less than 1% of VC firms had internal initiatives and experts dedicated to their digital transformation in 2023, we already see the first purist players pushing the limits even further: Quant VC firms. A model that takes the human out of the equation and exclusively relies on data and algorithms, just like Quant Funds did when disrupting the hedge fund industry in the 1990s.
I wrote about the three different models “handcraft VC”, “augmented VC”, and “quant VC” about a year ago here. Since then, I had several controversial discussions about the future of VC and the role of data & AI. While I’m personally a strong believer in an augmented VC approach for early-stage lead investors, I truly enjoy it when others challenge my perspective and have strong arguments to back up their convictions.
Today, I’m incredibly excited to have Guy Conway, the Co-Founder of Quant VC firm Koble.ai who will also join us as a speaker at the Data-driven VC Summit 2024 in early May, share his worldview and why he believes that pure quant strategies will win.
Thank you Guy for your thought-provoking write-up below🌶️🌶️
The AI Backlash Has Already Begun
In recent weeks, negative news headlines have started to proliferate. It seems the media can't make up its mind; journalists hypothesize that AI means death for our species, whilst also claiming that AI is a massive bubble. The schadenfreude from armchair skeptics and full-blown luddites is palpable. It’s almost as if people want us to fail.
What is true, is that the hype around artificial intelligence is normalizing.
According to data compiled by Bloomberg and Apollo chief economist Torsten Sløk, mentions of “AI”, “Machine Learning”, or “Generative AI” on earnings calls decreased from 517 instances in Q4 2023 to just 198 in Q1 2024.
This is healthy, and a net positive for the tech industry, since it reflects the maturation of AI and its wholesale adoption. Just as companies do not claim to be “Internet-driven”, they should not claim to be “AI-driven”, either.
Data Extremism
As the corporate narrative around AI is maturing, so too is the way investors think about integrating data and AI into their operations. The application of Data Science to the dark art of Venture Capital – and its impact on the investment process – is still nascent, and will take many years before its true impact can be measured and understood.
VCs have made decent progress in adopting tools to streamline operations and bring more rigor to the investment process, which remains stubbornly human. But how much further can Data-driven VCs go before hitting a ceiling in terms of how much internal spending can be allocated to refreshing the data stack?
And even if spending on data was unlimited, there is still the immovable (and seemingly ubiquitous) belief that data is the input; only humans can run the Investment Committee. How strange that we let AI discover new drugs for us, and then diagnose us with the illnesses with which those drugs should treat us, but the widely held industry belief remains that AI can’t tell us if a startup investment is the right one!
A growing subset of investors is calling for a more hardcore approach – one that excludes humans from the investment process entirely. But this camp is very much in the minority.
Full automation of the startup selection process requires technology that’s orders of magnitude more complex and capable than that which is currently accessible to mainstream VCs. But purist (some might say, extremist) “Quantitative VCs” are quietly building in the background.
Going Beyond the Consensus
AI is perfectly suited to helping humans manage informational overabundance, and deconstruct the Power Law to demystify the “miracle” of Venture Capital. Use cases are plentiful and well documented:
Deal flow filtering – analyse large datasets of startups, identifying patterns, trends, and KPIs and allowing fund managers to filter and prioritise investment opportunities.
Due diligence – conduct due diligence on startups by analysing vast amounts of unstructured data, such as news articles, social media posts, industry reports, and financial statements.
Predictive analytics – leverage data to develop predictive models that estimate the success or failure of startups.
Portfolio Management – manage portfolios more effectively by continuously monitoring and optimising the performance of existing investments.
Domain expertise – provide contextualised insights at the company and industry level, simulating and scaling the expertise of practitioners.
Taken as a whole, these are significant upgrades to the status quo. But Quant VCs argue they can go further, disrupting old-school investment methodologies and reliably hitting the upper echelons of the asset class.
What we've seen with Data-driven VC is about sourcing and discovery and picking off the low-hanging fruit of automating internal processes.
Quant VC goes much further – it’s about investment selection and portfolio construction too. In the long term, there’s a drive toward automation of the end-to-end capital deployment process.
The mean return in early-stage VC over 20 years is 21.3% IRR, yet median returns are at 5% IRR. No asset class (other than Crypto) can even get close to that. Adoption of Quant VC means we should consistently be able to hit the juicy returns the asset class offers. And that comes down to being better at picking the winners and the science of portfolio construction. Most VCs fail to do this.
History Repeating
Years ago, hedge funds disrupted public markets with data science. In the public markets, algorithmic trading decisions now represent up to 75%+ of total trading and AUM of $1+ trillion.
The history of the hedge fund industry can teach us much about the future of Venture Capital. Jim Simons – the billionaire founder of quantitative hedge fund Renaissance Technologies – is perhaps the greatest money-maker in modern financial history, with a track record for value accretion that surpasses Warren Buffett, George Soros and Ray Dalio.
A mathematician and code-breaker by trade, he decided to apply his esoteric expertise to financial markets by adopting a radical approach. When he started his fund, Simons didn't hire finance people; he hired physicists, mathematicians and computer scientists, tasking them with amassing mountains of data and building algorithms that could decode the hidden patterns that underlie public markets.
Buckminster Fuller said that:
“You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”
This is what Quant VCs are doing today. Just as Renaissance Technologies’ team of misfits “solved” public market investing, a low-profile community of programmers, AI researchers, and quants are re-engineering startup investing. In practice, “solving” Venture Capital means “building a new model”; using Data Science to understand and monetise the underlying network effects that drive value accrual in startups.
“The flagship fund of Renaissance Technologies, the first ever quant outfit, established in 1982, earned average annual returns of 66% for decades.” And just like the first wave of quant hedge funds, pioneering Quant VCs will make the juiciest returns.
Pushback
The Quant VC model is controversial and borderline divisive. A lot of very smart people object to the concept. They typically cite two major challenges:
1. How can an algorithm get allocation?
Investors believe the best deals are largely oversubscribed and super hot. But in reality, these are simply hyped deals. Instead, we should consider VC as two distinct asset classes; Pre-Seed & Seed, and Series A+.
It’s the Quant VC’s challenge to convince LPs that at Pre-Seed and Seed there isn't an access problem. In fact, 85%+ of deals at Pre-Seed have no recognisable institution attached to them.
Of course, there will be some deals that Quant VCs won't get access to (like every other fund), but at Pre-Seed and Seed there are more than enough high-quality deals that are not hyped that investors can play in.
Certainly, in the later stages of VC at Series A+ there is a big access problem. Quant VCs operating at this stage tend to optimize for price (assuming they have found a way to secure allocation).
2. How do Quant VCs add value post-investment? Will founders really want “hands-off” money on the cap table?
Quant VC resonates strongly with founders and investors, offering a compelling alternative to VC’s busted empathy model. By taking humans out of the equation, quant investors clarify expectations. No TED talk platitudes; no “partnerships”; no false empathy. Only algorithms – fast, quiet, hands-off, fair.
Not only are Quant VCs sitting on tons of data, but also they have the know-how to use it effectively. For example, they can use this data to provide amazing insights for founders to support their fundraising, hiring, market analysis, and competitor tracking.
Playing the Endgame
VC is a network business. One that is effectively capped by the scalability of human relationships. There is a cognitive gap in how many sectors and companies an individual investor can deeply understand without the help of data and technology. Technology offers a solution to this limitation, enabling investors to source and screen huge deal flow volumes.
Data-driven VCs integrate data points into human investment committees, whilst Quant VCs make investment decisions that are exclusively driven by algorithms. The difference lies not only in methodology but also in endgame. To fully understand what is happening, we must look beyond the “How” to the “Why”.
Quant VCs are not only looking to improve the investment process and the returns they can generate for LPs. They are looking to redefine and reposition the venture asset class.
Increasing the scale of startup analysis and investing can drive portfolio diversification and reduce volatility, generating risk-adjusted returns that significantly outperform industry averages. And by smoothing the return profile of Venture Capital, Quant VCs believe they can grow the asset class and route capital more efficiently to startups that truly deserve it.
That’s it for today. I hope you enjoyed this thought-provoking guest post and would love to hear your thoughts on the future of VC: traditional vs data-driven/augmented vs quant VC. What do you think?
Stay driven,
Andre
PS: We just announced the virtual Data-driven VC Summit 2024 on 5-8th May, limited early bird tickets here
Thank you for reading this episode. If you enjoyed it, leave a like or comment, and share it with your friends. If you don’t like it, you can always update your preferences to receive just the regular Thursday “Essays”, just the Tuesday “Insights”, or both. Subscribe below and follow me on LinkedIn or Twitter to never miss data-driven VC updates again.
What do you think about my weekly Newsletter? Love it | It's great | Good | Okay-ish | Stop it
If you have any suggestions, want me to feature an article, research, your tech stack or list a job, hit me up! I would love to include it in my next edition😎
Great to get this piece.
The interesting point is access. The proof of the pudding is in the investing.